PHYSICAL REVIEW B

VOLUME 62, NUMBER 9

1 SEPTEMBER 2000-I

Effect of Mn-site doping on the magnetotransport properties of the colossal magnetoresistance compound La2Õ3Ca1Õ3Mn1Àx A x O3 „AÄCo,Cr; xÏ0.1… F. Rivadulla* and M. A. Lo´pez-Quintela Physical Chemistry Department, University of Santiago de Compostela, E-15706 Santiago de Compostela, Spain

L. E. Hueso, P. Sande, and J. Rivas Applied Physics Department, University of Santiago de Compostela, E-15706 Santiago de Compostela, Spain

R. D. Sa´nchez Centro Ato´mico de Bariloche, 8400-San Carlos de Bariloche, Argentina 共Received 29 February 2000兲 In this paper we show the effect of Mn site substitution (⭐10%) by Co and Cr on the magnetotransport properties of ceramic samples of La2/3Ca1/3MnO3 . Resistivity, magnetization, and magnetoresistance were systematically investigated as a function of doping. An increase in resistivity and a diminution of metalinsulator transition and Curie temperatures was observed as a consequence of both Co and Cr doping. The evolution of the number of neighbors ferromagnetically coupled to Mn was studied in each sample, and the results suggest some degree of ferromagnetic coupling between Cr3⫹ and Mn ions. Implications of both ferromagnetic Cr 3⫹ -O-Mn3⫹ superexchange and double exchange interactions are discussed. We also found the existence of an optimum doping level that enhances colossal magnetoresistance. This has been explained considering electron localization due to magnetic disorder induced by doping in the Mn site.

I. INTRODUCTION

The discovery of colossal magnetoresistance 共CMR兲 in mixed valence manganites promoted an extensive number of works about these compounds in the last years.1,2 Probably one of the most studied is the archetypal La2/3Ca1/3MnO3 . 3 Schiffer et al.4 obtained the phase diagram of the whole series La1⫺x Cax MnO3 (0⬍x⬍1) showing a remarkable correlation of magnetic order and electric transport with x. The simultaneous occurrence of metallicity and ferromagnetism 共FM兲 for intermediate doping levels (0.2⭐x⭐0.45) has been explained by Zener in the framework of a ferromagnetic double exchange 共DE兲 interaction:5 the unpaired e g electron in the high spin configuration of Mn3⫹ moves to a neigboring Mn4⫹ , such a transfer being favored if the Mn4⫹ which 3 ) is to receive the electron has its own spins (S⫽3/2, t 2g 3⫹ aligned parallel with respect to Mn spins. In this scenario, electron delocalization between Mn ions leads to a lowering of energy, and will favor a ferromagnetic ordering of the magnetic moments. Several works have shown how DE interaction, and hence FM coupling and CMR, can be tuned by changing the mean ionic radius at the rare-earth site.6,7 But although this is one of the most studied topics of solid-state physics, not too much is known about the effect of substitution in the Mn site, even though it acts at the heart of DE interaction. Ahn et al.8 reported the effect of Fe doping on the magnetic and CMR properties of La0.63Ca0.37MnO3 . In this case, electron hopping between Fe and Mn is impeded by the lack of available states in the Fe e ↑g band. Consequently, Fe causes a depletion of the available hopping sites and DE is suppressed. Similar results were found by Blasco et al.9 and by Y. Sun et al.10 when doping La0.67Ca0.33MnO3 with Al3⫹ 0163-1829/2000/62共9兲/5678共7兲/$15.00

PRB 62

and Ga3⫹ , respectively. J. R. Sun et al.11 reported marked differences for Fe3⫹ and Ge4⫹ doping in La0.7Ca0.3MnO3 reflecting different effects due to electron or hole trapping. Ghosh et al.12 doped La0.67Ca0.33MnO3 with a 5% of the entire first transition element series and found a correlation between the maximum MR, the lattice parameter, and the ionic radii of these elements. Gayathri et al.13 studied the La0.7Ca0.3Mn1⫺x Cox O3 series and found that Co substitution suppresses the CMR characteristic of the undoped manganite over the entire temperature and composition range. They argue a transition from a long-range ferromagnetic order to a cluster glass-type ferromagnetism due to Co doping, even for the lowest degree of substitution. Moreover, for x⬎0.05 they observed semiconducting behavior in the whole temperature range. However, Rubinstein et al.14 found that up to 20% of Co substitution, La2/3Ca1/3MnO3 is metallic below the ferroparamagnetic transition temperature. On the other hand, it has been recently reported that Cr doping in Sm0.5Ca0.5MnO3 destroys the low-temperature antiferromagnetic-charge ordered state to induce a FM-metallic component in this material.15 This has been used as a proof of the ability of Cr to be DE coupled with Mn ions. The aim of this paper is to study the effect of a low degree of substitution by two different magnetic ions under controversy 共Co and Cr兲 on the magnetotransport properties of the CMR compound La2/3Ca1/3MnO3 . As we have checked by x-ray diffraction, no appreciable changes in the lattice structure are produced at these low doping levels, and the effects due to electronic mistmach between Mn and Co/Cr become the only relevant. II. EXPERIMENTAL DETAILS

Two series of polycrystalline La2/3Ca1/3Mn1⫺x Cox O3 共LCMCo兲 and La2/3Ca1/3Mn1⫺x Crx O3 共LCMCr兲 (x⫽0, 5678

©2000 The American Physical Society

EFFECT OF Mn-SITE DOPING ON THE . . .

PRB 62

5679

TABLE I. Theoretical and experimental stoichiometry, referred to Ca content, of ceramic samples obtained by x-ray fluorescence spectrometry. Theoretical

Experimental

% Co, Cr in Mn site

La0.67Ca0.33Mn La0.67Ca0.33Mn0.975Co0.025 La0.67Ca0.33Mn0.95Co0.05 La0.67Ca0.33Mn0.925Co0.075 La0.67Ca0.33Mn0.9Co0.1 La0.67Ca0.33Mn0.975Cr0.025 La0.67Ca0.33Mn0.95Cr0.05 La0.67Ca0.33Mn0.925Cr0.075 La0.67Ca0.33Mn0.9Cr0.1

La0.6698Ca0.33Mn1.017 La0.672Ca0.33Mn1.03Co0.0259 La0.655Ca0.33Mn0.988Co0.053 La0.655Ca0.33Mn0.95Co0.078 La0.663Ca0.33Mn0.92Co0.105 La0.65Ca0.33Mn1.00Cr0.024 La0.686Ca0.33Mn1.09Cr0.0735 La0.677Ca0.33Mn0.97Cr0.074 La0.684Ca0.33Mn0.95Cr0.10

2.45 5.09 7.59 10.24 2.34 6.79 7.09 9.52

0.025, 0.05, 0.075, 0.10兲 were prepared by standard ceramic procedures. Stoichiometric mixtures of La2 O3 , CaCO3 , Mn2 O3 , MnO2 , and Co(NO3 ) 2 or Cr(NO3 ) 3 共at least 99.99% in purity兲 were grounded, pressed, and heated in air at 1000 °C for 10 h. After this step, the Co/Cr nitrates has been confirmed to transform into their oxides by x-ray diffraction. After grinding, they were pressed into pellets and sintered in air at 1100 °C for 70 h, 1200 °C for 30 h, and 1300 °C for 100 h with intermediate grindings. Cationic content of these samples was measured by x-ray fluorescence with a Siemens SRS 3000 spectrometer. Samples were melted with BO4 Li, CO3 Na2 , NO3 NH4 , and LiBr in a Pt crucible at 1100 °C. More than 20 reference patterns with different concentrations of La, Ca, Mn, Cr, and Co were prepared in order to obtain the exact stoichiometry of ceramic samples. Structural information was derived using the program LS1, based on the Rietveld method, and fitting profiles to pseudo-Voigt functions.16 Zero-field-cooling 共ZFC兲 (M ZFC ) and field-cooling 共FC兲 (M FC ) magnetization curves were measured with a vibrating-sample magnetometer between 100 and 300 K at 10 Oe. Under such small applied field, the Curie temperature (T C ) can be obtained very accurately from the peak in dM FC /dT . The dc magnetic susceptibility ␹ (T) was measured above T C up to 1000 K with a Faraday balance at 5 kOe.

also synthesized under an O2 current in order to prevent for possible oxygen vacancies, but the results were practically identical to the sample made under air atmosphere. In any case, the volume variation with respect to the x⫽0 sample is lower than 0.5%. B. Magnetotransport properties

Figure 2 shows the temperature dependence of resistivity 共left panel兲 and magnetization 共right panel兲 of LCMCo 共top兲 and LCMCr 共bottom兲. Higher doping level results in lower metal-insulator transition temperatures (T M I ) and higher residual and peak resistivities. On the other hand, above T M I all the samples follow a universal curve characteristic of a thermally assisted conduction mechanism. The Curie temperature 共Fig. 3兲 also reduces with doping, following T M I .

III. EXPERIMENTAL RESULTS A. Structure and chemistry

In Table I we show the comparison between expected and experimentally determined stoichiometry in ceramic samples. Most of them present only slight deviations from the desired compositon. In the rest of the paper we will refer to the samples using the experimental compostion. Structural parameters deduced from the Rietveld refinement of the x-ray-diffraction patterns of ceramic samples are summarized in Fig. 1. All the compounds are orthorhombic 共Pbnm兲 with the O⬘ structure (c/ 冑2⬍a⬍b). Substitution with Co or Cr does not substantially change the structure of the undoped compound, although slight variations in the unit-cell volume (⬍0.5%兲 are evident due to differences in the ionic radii of Co/Cr and Mn. Sample La0.684Ca0.33Mn0.95Cr0.10O3⫾ ␦ does not follow the gradual reduction of volume observed in the series. This sample was

FIG. 1. Evolution of volume and lattice parameters with Co 共top兲 and Cr 共bottom兲 concentration extracted from Rietveld refinement of x-ray powder data of ceramic samples. For the adjustment we have used site occupancies obtained from x-ray fluorescence analysis.

5680

F. RIVADULLA et al.

PRB 62

FIG. 2. Temperature dependence of resistivity 共left panel兲 and magnetization 共right panel兲 of LCMCo 共top兲 and LCMCr 共bottom兲 series.

However, the linear reduction of T C produced by Co is not observed in Cr series, where a tendency to stabilization at about 210 K is clearly observed. Anyway, T C is always smaller than T M I , except in La0.684Ca0.33Mn0.95Cr0.10O3⫾ ␦ . This could be the signature of the presence of a high degree of oxygen vacancies in this sample. Between T C and T M I , samples are in a mixed state of paramagnetic 共PM兲 and ferromagnetic phases.17 The presence of a short-range FM order between T C and T M I produces a charge localization and an abrupt rise of resistivity, but the sample remains metalliclike (d ␳ /dT⬎0) if transport time is comparable to the lifetime of these short range FM correlations and some degree of delocalization persist. In Fig. 4共a兲 we show the magnetoresistance isotherms (T⫽T M I ) measured up to 7.5 kOe for some samples as a function of doping. In both series, LCMCo 共top兲 and LCMCr

共bottom兲 we observed a systematic increase of the % CMR up to ⬃5% of doping. For higher doping % CMR decreases again 关Fig. 4共b兲兴. This reflects the existence of an optimum doping level that favors CMR, whatever the nature of the dopant. It should be noted that we have doubled the value of the intrinsic magnetoresistance of La2/3Ca1/3MnO3 associated to the FM-PM transition with only ⬃5% of Co doping. In Fig. 5 M FC-ZFC curves, representative of the general behavior of the two series of samples are presented. For x⭓0.05, a strong dependence of M ZFC curves below T C becomes evident in Co samples. This effect is not observed in Cr samples, where a temperature-independent behavior of the M ZFC , similar to the x⫽0 sample, is observed in all the series. The drop of the M ZFC curve for x⭓0.05 is an indication of the absence of long-range FM order and is a signature of cluster-glass-like behavior in highly doped Co samples. On the other hand, the almost temperature-independent M ZFC curves of Cr samples below T C is a signature of a well established long-range FM order. High-temperature magnetic susceptiblity 关 ␹ (T) 兴 follows a Curie-Weiss law. Below ⬇1.4⌰ 共where ⌰ is the paramagnetic Curie-Weiss temperature兲 ␹ ⫺1 (T) shows a positive curvature and T C ⬍⌰. From the linear part of the plot ␹ ⫺1 (T) we have derived the Curie constant 共C兲 and ⌰ for each sample 共see Table II兲.

IV. DATA ANALYSIS AND DISCUSSION

FIG. 3. Dependence of T C with % Co (䉭) and % Cr (䊉). Lines are guides to the eye.

The observed reduction of T M I and T C 共Figs. 2 and 3兲 cannot be described by a variation in the Mn3⫹ /Mn4⫹ relationship, as the phase diagram for La0.67Ca0.33MnO3 does not justify the large temperature shifts observed for the low dop-

PRB 62

EFFECT OF Mn-SITE DOPING ON THE . . .

5681

FIG. 5. FC-ZFC curves representative of the general behavior in both series. Doping levels x⭓0.05 produces the observed temperature dependendence of the ZFC curve below T C in Co samples 共top兲. This is not observed in Cr series 共bottom兲.

FIG. 4. 共a兲 % CMR vs H measured at the metal-insulator transition temperature of each sample. 共b兲 CMR at 7.5 kOe for LCMCo (䉭) and LCMCr (䊉).

ing levels discussed here.4 On the other hand, doping does not substantially affect to the paramagnetic resistivity. At high temperatures, multiphonon-assisted hopping of small polarons occurs between sites through a thermally activated process. Transport measurements by Jaime et al.18 demonstrated that the conductivity in the PM regime of mixed valence manganites is dominated by the thermally activated hopping of small polarons in the adiabatic limit. We have obtained values for the polaron hopping energy E ␳ ⬇175 ⫾15 meV, irrespective of dopant and doping level. De Teresa et al.19 found that small variations in the Mn3⫹ content of CMR manganites have extraordinary effects in the polaron hopping energy, evidencing the Jahn-Teller 共JT兲 nature of these polarons. The almost constant value of E ␳ is an indication of the practical invariance in the Mn3⫹ /Mn4⫹ ratio in our samples. Very recently, Gosh et al.12 reported the

effect of ionic radii mistmach at the Mn site for all the elements of the first transition series at a fixed dopant concentration 共5%兲 in La0.7Ca0.3MnO3 . They attribute a central role to lattice strains in controlling the magnetotransport of the system. However, a variation of ⬃20 K for x⬇0.025 cannot be justified in the basis of their results. Moreover, the enhancement of % CMR 共measured at T M I ) with doping 共Fig. 4兲 is in contradiction with the strain-CMR correlation proposed in Ref. 12 since slightly lower % CMR should be expected for 5% of Co and Cr substitution. Since the Mn3⫹ /Mn4⫹ ratio is approximately constant for all the samples, the variations in ⌰ reflect the changes in the isotropic exchange interactions, and the degree of magnetic frustration induced with doping can be measured through the ratio T C /⌰. For x⫽0, T C /⌰⬇0.71, close to the value of 0.78 predicted by the constant coupling approximation 共CCA兲.20 This ratio tends to decrease with doping 共see Table II兲 but is more dramatic for Co than for Cr samples. This result suggests than the magnetic frustration induced by Co TABLE II. Values for the Curie constant, Curie-Weiss, and Curie temperatures obtained from susceptibility measurements. Compound La0.6698Ca0.33Mn1.017O3⫾ ␦ LCMCo 2.45% 5.09% 7.59% 10.24% LCMCr 2.34% 6.79% 7.09% 9.52%

C 共emu K/mol兲 ␪ 共K兲 T C 共K兲 T C / ␪ 2.8共1兲 2.7共1兲 2.5共1兲 2.7共2兲 2.6共1兲 2.7共1兲 2.6共2兲 2.6共1兲 2.6共3兲

368 343 392 353 371 383 342 332 339

260 240 209 184 170 240 220 217 208

0.706 0.7 0.53 0.52 0.43 0.627 0.643 0.654 0.614

5682

F. RIVADULLA et al.

FIG. 6. Evolution of the effective number of neighbors ferromagnetically coupled to Mn as a function of doping in LCMCo (䉭) and LCMCr (䊉). Lines are guides to the eye.

doping in the Mn sublattice is larger than induced by Cr. Within the CCA, the effective number of neighbors FM coupled to Mn can be derived from the values of T C /⌰. Results are plotted in Fig. 6. While Co produces a reduction approximately linear of the number of neighbors with doping, Cr only reduces the neighbors in a small quantity and tends to a stable value of about 5.25. From this result it follows that Co does not participate in DE interaction, but there is some degree of FM coupling between Cr and Mn ions. The electronic configuration of Co is very complicated due to the existence of several possible spin states. Apart from low and high spin, Bahadur et al.21 reported ferromagnetic resonance data supporting the existence of intermediate 5 1 4 1 e g ) and Co4⫹ (t 2g e g ) in spin configurations for Co3⫹ (t 2g mixed-valence cobaltites. Moreover, the possibility of temperature induced transitions from low to localizedintermediate and to delocalized-intermediate spin states has been demonstrated in the parent compound LaCoO3 . 22 On 3 0 e g ) is the only ion replacing manthe other hand, Cr3⫹ (t 2g ganese in the LCMCr series. In the view of these electronic configurations, it seems reasonable to think in Cr3⫹ as the only possible candidate to establish a FM interaction with Mn ions. Cr3⫹ is isoelectronic with Mn4⫹ and DE interaction should be possible between Mn3⫹ and Cr3⫹ , although the different effect of the crystal field over Mn4⫹ and Cr3⫹ must cause energetic differences between e ↑g levels of Mn3⫹ and Cr3⫹ strong enough to make electronic exchange between them not so effective as between Mn3⫹ and Mn4⫹ . Moreover, Cr3⫹ ions take their positions presumably at random at the time of sample preparation, whereas only electron transfer is necessary to produce ordering of Mn3⫹ and Mn4⫹ ions. On the other hand, Cr3⫹ -O-Mn3⫹ superexchange interaction is ferromagnetic, and should be taken into account in order to explain our results.23 This interaction leads to FM allignement of Cr3⫹ and Mn3⫹ keeping the number of neighbors FM coupled to Mn near to 6. In a neutron-diffraction study of the magnetic structure of La(Mn,Cr)O3 , Bents24 showed that a low percentage of Cr3⫹ destroys the A-type AF structure of LaMnO3 , and produces the spin reversal of some of its nearest Mn3⫹ neighbors via superexchange coupling leading to a FM moment. The rate of destruction of the AF phase by Cr3⫹ -O-Mn3⫹ FM superexchange is compa-

PRB 62

rable to that observed when Mn 3⫹ is replaced by Mn4⫹ and FM DE interaction is established. Very recently, a number of experimental works built up the idea of the participation of Cr3⫹ in DE. In example, paramagnetic-insulating to ferromagnetic-metallic transitions were induced in Sm0.5Ca0.5MnO3 by Cr doping (0.05⭐x ⭐0.09). 15 Raveau et al. 25 find that Cr, Co, and Ni kills the charge ordering state in Pr0.5Ca0.5MnO3 restoring a metalinsulator transition high values of magnetization. However, the maximum value of the magnetic moment is larger for Cr doped samples, and the doping effect is extended for a largest range of Cr substitution 共3–10%兲. As a result of this doping effect, participation of Cr in de DE mechanism has been proposed. However, Damay et al.26 reported an induced antiferromagnetic-insulating to ferromagnetic-metallic transition in the charge ordered 共CO兲 compound Pr0.6Ca0.4MnO3 by doping with Fe3⫹ (3d 5 ), Al3⫹ (3d 0 ), Ga3⫹ (3d 0 ), and Mg2⫹ (3s 0 ). A strong FM component (2.1␮ B for x ⫽0.035) is induced by iron doping. Since these ions cannot establish an effective DE interaction with Mn due to their electronic configuration, it easily follows that the weakening of the CO-AF ground state by Mn site doping in CO manganites is developed by the introduction of some degree of disorder in the Mn sublattice. On the other hand, the increasing magnetic moment induced by Cr doping can be successfully explained considering FM superexchange interaction only. The presence of this interaction also justify why this doping induced FM component is restricted to a small percentage of substitution. For higher percentages of Cr3⫹ , formation of Cr3⫹ -O-Cr3⫹ and Cr3⫹ -O-Mn4⫹ pairs become more favorable, leading to growing AF superexchange interactions. So, we believe that participation of Cr in the DE interaction cannot be deduced from these results, and the behavior observed in CO manganites cannot be extrapolated to other members of the series, like La2/3Ca1/3 MnO3 . Although in this work we clearly show the existence of FM coupling between Mn3⫹ and Cr3⫹ in La2/3Ca1/3MnO3 , we cannot ensure that Cr3⫹ participates in the DE, as we believe that the possiblity of FM superexchange interaction must been taken into account in order to explain these results. We showed how Co makes the Mn sublattice more disordered that Cr, from a magnetic point of view. This magnetic disorder leads to a localization of conduction electrons, in a similar way as it was pointed out by Anderson 共Anderson localization兲.27 In the Anderson model a mobility edge E b can be defined to separate localized states of the tail band from extended states of the middle of the band. Metalinsulator transition occurs when the Fermi level (E F ) moves, through the mobility edge, from a extended state to a localized state. The increasing residual resistivity and decreasing T M I observed with doping in our samples is a consequence of increasing disorder. Moreover, replacement of Mn by Co or Cr leads to some degree of magnetic disorder superimposed to the nonmagnetic one 共see Fig. 6兲. Doping produces an effect very similar to the effect of temperature, i.e., E b grows and becomes larger that E F , which now is a localized state. As CMR isotherms of Fig. 4共a兲 are measured at the corresponding T M I , the effective temperature is the same for each sample. The effect of the applied field is to alter the localization length, reducing E b , being E F in the middle of

PRB 62

EFFECT OF Mn-SITE DOPING ON THE . . .

the extended states. For doping levels larger than ⬃5%, E b ⰇE F and the effect of 7.5 kOe is not so strong as to make E b ⬍E F . Sheng et al.28 argued that the presence of a certain quantity of randomness, along with the effective hopping dissorder of the DE model, produces an Anderson metal-toinsulator transition associated to the FM-PM transition. These authors proposed the existence of an optimum value of nonmagnetic dissorder strength that favors the CMR effect. Our data fully support their theoretical predictions, although the type of disorder here introduced is not exactly the same referred by Sheng. As we have mentioned, nonintrinsic effects appear for x ⭓0.05 in Co samples. The number of DE-inactive species introduced by Co doping at the Mn site, is high enough to break the long-range magnetic coherence when x⭓0.05. Ritter et al.29 proposed that ⬃5% of Mn site vacancies are responsible for the appearance of a cluster-glass state in La3/(3⫹ ␦ ) Mn3/(3⫹ ␦ ) O3 共% Mn4⫹ ⫽2 ␦ ⫻100). The assumption of the presence of high magnetic disorder above x ⬇0.05 in Co samples is supported by the behavior of the M FC-ZFC curves as a function of doping 共Fig. 5兲. M ZFC for x⬍0.05 are almost independent of T below T C as in the pure manganite (x⫽0). This is a signature of a well established long-range FM order. However, for x⭓0.05 a strong dependence of M ZFC curves below T C becomes evident in Co series. The drop of the M ZFC curve for % Co⫽5.09 is an indication of the absence of long-range FM order and is a signature of cluster-glass-like behavior. From these results we conclude that the most likely state in our highly doped Co samples is formed by FM clusters in a background with ferro-antiferromagnetic competing interactions, which probably lead to an overall cluster glass behavior.

*Corresponding author. Email address: [email protected] 1

R. von Helmolt, J. Wecker, B. Holzapfel, L. Schultz, and K. Samwer, Phys. Rev. Lett. 71, 2331 共1993兲. 2 S. Jin, T. H. Tiefel, M. McCormack, R. A. Fastnacht, R. Ramesh, and L. H. Chen, Science 264, 413 共1994兲. 3 G. H. Jonker and J. H. van Santen, Physica 共Amsterdam兲 16, 337 共1950兲. 4 P. Schiffer, A. P. Ramirez, W. Bao, and S. W. Cheong, Phys. Rev. Lett. 75, 3336 共1995兲. 5 C. Zener, Phys. Rev. 82, 403 共1951兲. 6 H. Y. Hwang, S.-W. Cheong, P. G. Radaelli, M. Marezio, and B. Batlogg, Phys. Rev. Lett. 75, 914 共1995兲. 7 ˜ ol, J. L. GarciaJ. Fontcuberta, B. Martinez, A. Seffar, S. Pin ˜ oz, and X. Obradors, Phys. Rev. Lett. 76, 1122 共1996兲. Mun 8 K. H. Ahn, X. W. Wu, K. Liu, and C. L. Chien, Phys. Rev. B 54, 15 299 共1996兲. 9 J. Blasco, J. Garcı´a, J. M. de Teresa, M. R. Ibarra, J. Pe´rez, P. A. Algarabel, C. Marquina, and C. Ritter, Phys. Rev. B 55, 8905 共1997兲. 10 Y. Sun, X. Xu, L. Zheng, and Y. Zhang, Phys. Rev. B 60, 12 317 共1999兲. 11 J. R. Sun, G. H. Rao, B. G. Shen, and H. K. Wong, Appl. Phys. Lett. 73, 2998 共1998兲. 12 K. Gosh, S. B. Ogale, R. Ramesh, R. L. Greene, T. Venkatesan, K. M. Gapchup, R. Bathe, and S. I. Patil, Phys. Rev. B 59, 533 共1999兲.

5683 V. CONCLUSIONS

We have tested the effect of Co and Cr doping on the magnetotransport behavior of the CMR compound La2/3Ca1/3MnO3 . We have found that Cr doping only produces a slight reduction in the number of neighbors ferromagnetically coupled to Mn, while Co dramatically reduces it. From these results it follows that Cr3⫹ is able to be FM coupled with Mn 3⫹ but success was achieved in accounting for these data on the basis of a FM Cr3⫹ -O-Mn3⫹ superexchange. It is then concluded that participation of Cr3⫹ in DE mechanism cannot be deduced from these experimental results. Moreover, our results show that both structural and magnetic disorder induced in the periodic lattice potential at low doping levels can be exploited to enhance the intrinsic CMR of the system, in perfect accordance with theoretical predictions. Mn-site doping of CMR compounds with T C over room temperature could be used to achieve higher magnetoresistance values at room temperature.

ACKNOWLEDGMENTS

We want to express our gratitude to Dr. M. T. Causa from the Centro Ato´mico de Bariloche 共Argentina兲 for her helpful discussion of the results and critical reading of the manuscript, Jorge Millos from the University of Vigo 共Spain兲 for compositional analysis of ceramic samples, and to M. Va´squez-Mansilla, from the Centro Ato´mico de Bariloche for measurements with the Faraday balance. Two of us 共F.R. and L.E.H.兲 want to acknowledge support from USC and MEC of Spain. We also thank MEC from Spain for financial support through Project No. MAT98-0416.

13

N. Gayathri, A. K. Raychaudhuri, S. Tiwary, R. Gundakaram, A. Arulraj, and C. N. R. Rao, Phys. Rev. B 56, 1345 共1997兲. 14 M. Rubinstein, D. J. Gillespie, J. E. Snyder, and T. M. Tritt, Phys. Rev. B 56, 5412 共1997兲. 15 A. Barnabe´, A. Maignan, M. Hervieu, F. Damay, C. Martin, and B. Raveau, Appl. Phys. Lett. 71, 3907 共1997兲. 16 L. Lutterotti, P. Scardi, and P. Maistrelli, J. Appl. Crystallogr. 25, 459 共1992兲; see also, for example, The Rietveld Method, edited by R. A. Young, IUCr-Monographs on Crystallography, Vol. 5 共Oxford Universtiy Press, New York, 1995兲, Chap. 8. 17 Z. W. Lin, J. W. Cochrane, G. J. Russell, X. L. Wang, S. X. Dou, and H. K. Liu, Appl. Phys. Lett. 74, 3014 共1999兲. 18 M. Jaime, H. T. Hardner, M. B. Salomon, M. Rubinstein, P. Dorsey, and D. Emin, Phys. Rev. Lett. 78, 951 共1997兲. 19 J. M. De Teresa, K. Dorr, K. H. Muller, L. Schultz, and R. I. Chakalova, Phys. Rev. B 58, 5928 共1998兲. 20 J. S. Smart, in Effective Field Theories of Magnetism 共W. B. Saunders Company, Philadelphia & London, 1966兲, Chap. V; M. T. Causa, M. Tovar, A. Caneiro, F. Prado; G. Ibanez, C. A. Ramos, A. Butera, B. Alascio, X. Obradors, S. Pinol, F. Rivadulla, C. Vazquez-Vazquez, M. A. Lopez-Quintela, J. Rivas, Y. Tokura, and S. B. Oseroff, Phys. Rev. B 58, 3233 共1998兲. 21 D. Bahadur, S. Kollali, C. N. R. Rao, M. J. Patni, and C. M. Srivastava, J. Phys. Chem. Solids 40, 981 共1979兲. 22 ˜ aris-Rodriguez and J. B. Goodenough, J. Solid State M. A. Sen Chem. 116, 224 共1995兲.

5684 23

F. RIVADULLA et al.

J. B. Goodenough and J. M. Longo, in Magnetic and Other Properties of Oxides and Related Compounds, edited by K.-H. Hellwege and A. M. Hellwege, Landolt-Bo¨rnstein, New Series, Group III, Vol. 4, Pt. a 共Springer, Berlin, 1970兲. 24 U. H. Bents, Phys. Rev. 106, 225 共1957兲. 25 B. Raveau, A. Maignan, C. Martin, and M. Hervieu, in Colossal Magnetoresistance, Charge Ordering and Related Properties of Manganese Oxides, edited by by C. N. R. Rao and B. Raveau 共World Scientific, Singapore, 1998兲, Chap. II.

26

PRB 62

F. Damay, A. Maignan, C. Martin, and B. Raveau, J. Appl. Phys. 82, 1485 共1997兲. 27 S. Elliot, in The Physics and Chemistry of Solids 共John Wiley & Sons, Chichester, UK, 1998兲, Chap. VI. 28 L. Sheng, D. y. Xing, D. N. Sheng, and C. S. Ting, Phys. Rev. Lett. 79, 1710 共1997兲. 29 C. Ritter, M. R. Ibarra, J. M. De Teresa, P. A. Algarabel, C. Marquina, J. Blasco, J. Garcia, S. Oseroff, and S. W. Cheong, Phys. Rev. B 56, 8902 共1997兲.